Questions regarding uncertainty in temperature measurement

[MRN370]

Hello,

I have some questions regarding the uncertainty in temperature measurements.
I use T-type thermocouple to meaure temperature. The thermocouple is connected to a NI 9214 datalogger with a default accuracy of 0.37 deg Celsius specified in the datasheet of the instrument. My first question is:

FIRST QUESTION: By looking at Table 3 of the datasheet, there are some conditions to select the accuracy of measurements. The measured temperature is noted on the first column, and a range of temperature is mentioned for the typical and maximum conditions on the other columns. Could you please let me know from which column (typical or maximum) in Table 3, I should take the accuracy according to the range of temperature measurements in my case specified above?

I did the calibration process by measuring and comparing the data of a precision thermometer with the thermocouple connected to the datalogger, all inside a thermal bath. The calibration is done for a range of -10 to 60 deg Celsius and the fitting polynomial is obtained by comparing the data from the precision thermometer and the thermocouple connected to the datalogger.

Now, in order to have the correct value of temperature read by the datalogger, I apply the polynomial to the data collected by the datalogger.

In order to calculate the uncertainty of measurements, I consider the systematic and random sources of errors. The systematic source includes the uncertainty due to the thermal bath, precision thermometer, and polynomial fitting.
My second question is this: ​

SECOND QUESTION: Do I need to include the uncertainty of the datalogger in the systematic error as well? Even after having a calibration for the thermocouple connected to the logger?

OR
It is already included in the calibrated data by fitting the polynomial?

In case that I need to include the uncertainty of the datalogger in the calculation, I will end up having uncertainty equal to 0.4 deg Celcius. However, I need to reach the accuracy of 0.1 deg Celcius. Then, the third question is:

THIRD QUESTION: Is there any way to reduce the uncertainty of the datalogger to be able to reach the 0.1 deg Celcius accuracy of temperature measurement?

I have attached a short description of my calculation. Please let me know if it is correct.

Thanks for your answer.

Mohammad